• No se han encontrado resultados

Research frontiers in null model analysis

N/A
N/A
Protected

Academic year: 2018

Share "Research frontiers in null model analysis"

Copied!
7
0
0

Texto completo

(1)

© 2001 Blackwell Science Ltd. http://www.blackwell-science.com/geb

337

E C O L O G I C A L S O U N D I N G

Global Ecology & Biogeography (2001) 10, 337– 343

Blackwell Science, Ltd

Research frontiers in null model analysis

NICHOLAS J. GOTELLI Department of Biology, University of Vermont, Burlington, Vermont 05405 U.S.A. E-mail: ngotelli@zoo.uvm.edu

ABSTRACT

Null models are pattern-generating models that deliberately exclude a mechanism of interest, and allow for randomization tests of ecological and biogeographic data. Although they have had a con-troversial history, null models are widely used as statistical tools by ecologists and biogeographers. Three active research fronts in null model analysis include biodiversity measures, species co-occurrence patterns, and macroecology. In the analysis of bio-diversity, ecologists have used random sampling procedures such as rarefaction to adjust for differ-ences in abundance and sampling effort. In the analysis of species co-occurrence and assembly rules,

null models have been used to detect the signa-ture of species interactions. However, controversy persists over the details of computer algorithms used for randomizing presence–absence matrices. Finally, in the newly emerging discipline of macro-ecology, null models can be used to identify con-straining boundaries in bivariate scatterplots of variables such as body size, range size, and popula-tion density. Null models provide specificity and flexibility in data analysis that is often not possible with conventional statistical tests.

Key words assembly rules, biodiversity, species co-occurrence, macroecology, null hypothesis, null model, rarefaction, taxonomic ratios.

When you have a new hammer, everything looks

like a nail

Old Russian proverb

Null models are statistical tests widely used in ecology and biogeography ( Nitecki & Hoffman, 1987; Manly, 1991; Gotelli & Graves, 1996). They have had a long and controversial history, but they continue to be a useful tool for describing pat-terns in ecological and biogeographic data (e.g. Colwell & Lees, 2000), particularly when conven-tional statistical analyses fall short. In this essay, I discuss three active research fronts where null models are (or should be) figuring prominently.

DEFINING THE NULL MODEL

A null model is ‘a pattern-generating model that is based on randomization of ecological data or random sampling from a known or imagined dis-tribution’ (Gotelli & Graves, 1996). In contrast to most other modelling approaches (Caswell, 1988),

the null model strategy is to construct a model that deliberately excludes a mechanism being tested. We want to know how well the data can be fitted by such a model (Hilborn & Mangel, 1997) — in other words, can the patterns in the real data be reproduced in a simple model that does not incorporate biologically important mechanisms? Or, do the data appear non-random with respect to the null hypothesis? If so, the analysis pro-vides some evidence in support of the mechanism (although it never can be taken as a ‘proof ’ of the mechanism in a strict Popperian framework).

Null models are rarely theoretical abstractions (e.g. Colwell & Hurtt, 1994); they usually begin and end with a real data set. To build a null model, an index of community structure, such as the amount of niche overlap ( Winemiller & Pianka, 1990), is first measured for the real data. Next, a null community is generated according to an algo-rithm or set of rules for randomization, and this same index is measured for the null community. A large number of null communities (typically 1000

(2)

338

N. J. Gotelli

or more) are used to generate a frequency histo-gram of index values expected if the null hypo-thesis is true. The position of the observed index in the tails of this null distribution is then used to assign a probability value to the pattern ( Manly, 1991), just as in a conventional statistical analysis.

In this essay, I am focusing on randomization tests that involve reshuffling, bootstrapping, or resampling from an observed data set, and Monte Carlo techniques that involve sampling from a specified parent distribution (Manly, 1991). Such tests can be tailored to address specific hypotheses about pattern, and to reflect sampling limitations and the spatial scale at which the data were col-lected. For these sorts of null models, many of the innovations have come from ecologists and bio-geographers, rather than from statisticians.

Constructing and testing a null model is straight-forward in theory and closely resembles hypo-thesis testing in conventional statistical analysis. Nevertheless, null models were highly controver-sial in the 1970s and 1980s (see reviews in Harvey

et al., 1983; Wiens, 1989; Gotelli & Graves, 1996;

Weiher & Keddy, 1999). Extended, and sometimes acrimonious, exchanges in the literature precipitated around philosophical and statistical issues.

The philosophical issues boiled down to whether the strict Popperian framework of falsification and parsimony (Popper, 1959) should be adopted by ecologists and biogeographers. An important subtext was a debate over the then-reigning com-petitive paradigm in ecology (Wiens, 1977; den Boer, 1981; Strong et al., 1984). These issues have largely subsided, even though the role of competi-tion in structuring the assembly of communities is still unresolved (Gotelli, 1999). As with cladistics, null models are now viewed as a useful tool for revealing pattern (or the lack thereof ) in natural communities, rather than as a Zeitgeist of the 1970s and 1980s debates over competition theory. The statistical issues in those debates centred around the precise details of null model construc-tion. Did certain algorithms inadvertently ‘smuggle in’ the effects they were designed to reveal (the ‘Narcissus Effect’ of Colwell & Winkler, 1984)? Conversely, were some algorithms predisposed to rejecting the null hypothesis for data sets that were entirely random (the ‘Jack Horner Effect’ of Wilson, 1995)? Using drier, more conventional labels, these are really controversies over whether

particular null model tests are predisposed to Type II error — incorrectly accepting a false null hypothesis (the Narcissus Effect), or Type I error — incorrectly rejecting a true null hypothesis (the Jack Horner Effect). This is an active area of research, and recent studies have quantified the frequencies of Type I and Type II errors by testing null model algorithms with random and structured data sets (e.g. Losos et al., 1989; Garvey

et al., 1998; Shenk et al., 1998; Gotelli, 2000).

Although the point has been made elsewhere (Connor & Simberloff, 1986), it bears repeating that the null hypothesis is not that communities are entirely random or have no structure (Roughgarden, 1983). Rather, it is that community structure is random with respect to the mechanism being tested. The null model can include as much struc-ture as is warranted by the data and the biology, as long as the mechanism of interest can be carefully excluded from the randomization. In practice, most null models are fairly simple in their randomization structure, if only because the kinds of data and biological information needed to construct more sophisticated null models (e.g. Graves & Gotelli, 1983) are usually lacking.

TAXONOMIC RATIOS AND THE MEASUREMENT OF BIODIVERSITY

Null models got their start in biogeography in the analysis of the species/genus and other such taxonomic ratios, which have long been used to describe community patterns and to infer levels of competitive interactions (reviews in Simberloff, 1970; Järvinen, 1982). A low species/genus ratio was interpreted as a product of strong intrageneric competition (Elton, 1946), which might limit con-generic coexistence (Darwin, 1859). Consistent with this hypothesis was the widespread observation that species/genus ratios were usually smaller for island than mainland communities (Elton, 1946). But what was the correct null hypothesis? In other words, what is the expected species/genus ratio in a community that is not structured by competition? The answer can be determined by establishing species/genus ratios for random communities that differ only in the total number of species they contain.

Such a null model was first explored for plant communities by Maillefer (1929), who used draws of species from a deck of shuffled cards to calculate

(3)

Null model analysis

339

the expected generic richness in small commu-nities. For animal communities, Williams (1947, 1964) elucidated these same patterns using statist-ical models and computer simulations. Although their work was ignored by ecologists for several decades (Järvinen, 1982), re-analyses of species/ genus ratios now suggest that island commu-nities harbour slightly more species per genus than expected by chance, in spite of the lower absolute number of species per genus expected in smaller samples (Simberloff, 1970). This find-ing is the opposite of what competition theory predicts, perhaps reflecting instead the similar dispersal potential and ecological requirements of congeneric species (the ‘Icarus Effect’ of Colwell & Winkler, 1984).

In a parallel, but independent effort, ecologists developed rarefaction as a method for comparing species richness of samples that differ in the num-ber of individuals collected (Sanders, 1968; Hurlnum-bert, 1971; Heck et al., 1975; Tipper, 1979). Although the motivation was quite different, the statistical problem is the same — how meaningfully to com-pare category/subcategory taxonomic ratios such as individuals/species or species/genus. Statistical methods (and software) have now been well developed to allow for meaningful comparisons that control for differences in both sampling effort and abund-ance (Gotelli & Colwell, in press).

Nevertheless, such well-understood sampling effects are frequently ignored in biodiversity studies. Ecologists have repeatedly made the error of com-paring richness per quadrat (species density) among sites differing in overall plant or animal dens-ity (e.g. Hubbell et al., 1999). These comparisons have confounded or equated differences in density with the differences in disturbance, successional, or productivity regimes that are being compared (Chazdon et al., 1999; Stevens & Carson, 1999). Even in small-scale experimental studies in which sampling effort is carefully controlled, standard-ization on the basis of number of individuals collected can completely change diversity patterns in experimental treatments ( McCabe & Gotelli, 2000). Species density (standardized by effort or area) and species richness (standardized by the number of individuals) may yield fundamentally different patterns of biodiversity; it is not yet clear under which circumstances each of these measures is most appropriate. Although rarefaction repres-ents the most basic (and least controversial) null

model, it needs to be more widely applied in bio-diversity studies at all spatial and temporal scales (Gotelli & Colwell, in press).

SPECIES CO-OCCURRENCE AND COMMUNITY ASSEMBLY RULES

No other issue in the history of ecology has been as contentious as the analysis of species co-occurrence patterns. Diamond’s (1975) monograph on the birds of the Bismarck Archipelago popular-ized the idea of using ‘assembly rules’ based on competitive interactions to interpret species co-occurrence patterns. In many ways, this monograph represents the apogee of the competition para-digm that was so heavily influenced by the work of G.E. Hutchinson and R.H. MacArthur.

In a sharply worded response, Connor & Simberloff (1979) attacked the logic of Diamond’s (1975) assembly rules and presented a pioneering null model analysis to determine the expected pattern of species co-occurrence in the absence of competitive interactions. This pair of papers sparked off a debate over the statistical analysis of presence– absence matrices that has continued unabated for over 25 years (e.g. Diamond & Gilpin, 1982; Gilpin & Diamond, 1982; Connor & Simberloff, 1983; Wilson, 1987; Roberts & Stone, 1990; Stone & Roberts, 1990; Manly, 1995; Sanderson et al., 1998; Gotelli & Entsminger, in press; Zaman & Simberloff, in press). For many ecologists, these baroque exchanges are synonymous with null models. No wonder that many have been wary to adopt null models for their own analyses!

Although the debate has become more focused, controversy persists. Connor & Simberloff (1979) introduced an important null model that random-izes presence–absence matrices, but maintains a fixed set of row (= species occurrences) and column (= island species number) totals. It has been a computational challenge to develop an algo-rithm that samples such matrices randomly and equi-probably (Roberts & Stone, 1990; Manly, 1995; Sanderson et al., 1998; Gotelli & Entsminger, in press; Zaman & Simberloff, in press). Unfortunately, some authors have made sweeping, premature claims about the validity of their methods, sometimes based on the analysis of a single data matrix. To evaluate these methods properly, the algorithms have to be tested with artificial data sets of two kinds: random data sets that contain no pattern,

(4)

340

N. J. Gotelli

and structured data sets that contain a mixture of structure and random noise (Gotelli et al., 1997). In this way it is possible to quantify the tendency towards Type I and Type II error in null model ana-lysis (Roxburgh & Matsuki, 1999; Gotelli, 2000). Such analyses are tedious and not very exciting, but they are essential if null models are to have a solid statistical foundation.

Other innovations in the study of species co-occurrence patterns include: the development of the ‘favoured states’ assembly rule for functional groups (Fox, 1987; Fox & Kirkland, 1992; Fox & Brown, 1993; Fox, 1999), incorporation of intra-specific spatial patterns into null models (Wilson & Gitay, 1995; Roxburgh & Chesson, 1998; Wilson, 1999), and development of explicit biogeographic criteria for designating realistic species source pools (Graves & Gotelli, 1993; Penev, 1997). In spite of continued interest in assembly rules and species co-occurrence (Weiher & Keddy, 1999), it is still unclear how general Diamond’s (1975) model is because so few data sets have been analysed with comparable methods. A recent meta-analysis (Gotelli & McCabe, in press) suggests that most communities exhibit non-random co-occurrence patterns that are indeed consistent with Diamond’s (1975) model, although there are intriguing differ-ences among taxonomic groups in the degree of non-randomness.

MACROECOLOGY

Seminal papers by Brown & Maurer (1986, 1987, 1989; see also Damuth, 1981) introduced the study of ‘macroecology’ (Brown, 1995; Blackburn & Gaston, 1998; Maurer, 1999), which emphasizes that the way species partition energy in nature is a fundamental constraint on community struc-ture (Brown, 1995). In many studies of macro-ecology, species-level variables such as body size, population size, and geographical range area are plotted in two-dimensional scatter plots in which each species represents a single data point. The plots often exhibit unusual polygonal shapes of points. However, to my eyes, they sometimes resemble the scattershot of bullet holes in rural road signs.

In any case, the next step is to draw in ‘bound-aries’ in the two-dimensional space, beyond which real data points never occur (absolute boundaries) or rarely occur (probabilistic boundaries).

Con-ventional statistics and regression analyses are of little help here because they merely pass a line through the centre of a cloud of points and (typic-ally) test for a slope of zero. But even if the slope of the regression is zero, the relationship between

x and y may change near the edge vs. the centre of the distribution.

For example, the plot of population density vs. body size of North American birds shows a pronounced peak (Brown & Maurer, 1987), per-haps suggesting an optimal intermediate body size. However, null model analysis suggests that this pattern may not be so unusual (Blackburn et al., 1990). The problem is that there are more species of intermediate body size, so the range of popula-tion sizes will tend to be greatest at intermediate body sizes. Thus, a peak may appear in the plot even if there is no constraint on the relationship between body size and population size.

Recently, sophisticated statistical methods have been proposed to analyse such data sets, includ-ing polynomial regression (Blackburn et al., 1990), quantile regression (Blackburn et al., 1992; Scharf

et al., 1998; Cade et al., 1999), path analysis (Thomson

et al., 1996), and other techniques (Garvey et al.,

1998). However, a null model approach may be even more effective (Blackburn et al., 1990). First, null models are often much simpler and easier to understand than some of these statistical models. Second, many of these statistical models require some arbitrary decisions to implement (e.g. which quantile should be analysed, or which path struc-tures should be compared and by what criterion?). Moreover, some of these tests require symmetrical, Gaussian data, which are often not found with macroecological variables, even after data trans-formation. Finally, formulating and testing the null model lays bare two critical questions that may be hidden or implicit in sophisticated statistical models.

The first question is how can boundaries in two-dimensional plots be established objectively? One promising approach is to use simple poly-gons defined by midpoints, medians, and ranges of observed data distributions (Gotelli & Entsminger, 2000). These polygons define triangular and pyr-amid shapes, but they do not depend on subjective or visual placement of potential boundaries.

Second, and more important, what is the appro-priate null distribution of points in bivariate space against which patterns should be compared? The

(5)

Null model analysis

341

unstated null hypothesis in many macroecological studies is that the distribution of points is isotropic, with an even density of species throughout the bivariate space; thus, the marginal distributions for each variable would be uniform. The boundaries of that space can be defined by the endpoints of the data, by continental limits on area, or by phylo-genetic extremes observed within a well-defined clade (Kochmer & Handel, 1986).

A different approach, taken by Enquist et al. (1995), is to use the marginal distributions of the x and y variables themselves to set the limits. As in classic randomization tests, the x and y

data are reshuffled, so the null model tests for unusual patterns in the covariance of x and y

(Garvey et al., 1998). Thus, the observation that few large-bodied species have large population sizes might not be unusual. If large-bodied cies are typically rare ( Van Valen, 1973) and spe-cies with large population sizes are uncommon (Preston, 1962), we should not be surprised to find that this combination of traits is rarely found together.

These two kinds of null distributions (uniform vs. marginals) tend to give different answers. Many data sets do not appear to be isotropic, so they show significant patterns when tested against the uniform null hypothesis. However, these same data sets may not exhibit unusual covariance struc-ture, so they appear random when tested against the marginals null hypothesis. Such contrasting results force investigators to consider expli-citly how assemblages would be structured in the absence of any biological constraints. Looking to the future, other null distributions are possible. Recent synthetic work on allometric scaling and constraints on body size (West et al., 1997, 1999a, 1999b; Enquist et al., 1998, 1999) may prompt a different sort of null distribution for macro-ecological studies.

In closing, I will issue a plea for plurality, and for researchers to eschew the ‘single hammer’ mentality. There is no one ‘correct’ method for constructing and interpreting a null model! The advantage of null models is that they provide flex-ibility and specificity that cannot often be obtained with conventional statistical analyses. Ironically, ecologists in the 1980s became polarized over null models and argued about a narrow subset of algorithms developed for testing patterns of com-petitive structure. Ecologists and biogeographers

should now exercise their imaginations and develop appropriate null models for addressing other research questions.

ACKNOWLEDGMENTS

I thank J.H. Brown, R.K. Colwell, M.V. Lomolino, R.J. Whittaker and an anonymous referee for comments on this essay. My ideas on null models benefited from conversations with the NCEAS Working Group in Biogeography and the Univer-sity of Vermont Ecology Discussion Group. My work on null models is supported by NSF grants BIR-9612109 and DBI 9725930.

REFERENCES

Blackburn, T.M. & Gaston, K.J. (1998) Some methodo-logical issues in macroecology. American Naturalist,

151, 68 – 83.

Blackburn, T.M., Harvey, P.H. & Pagel, M.D. (1990) Species number, population density and body size relationships in natural communities. Journal of

Animal Ecology, 59, 335 –345.

Blackburn, T.M., Lawton, J.H. & Perry, J.N. (1992) A method of estimating the slope of upper bounds of plots of body size and abundance in natural animal assemblages. Oikos, 65, 107 –112. den Boer, P.J. (1981) On the survival of populations

in a heterogeneous and variable environment.

Oecologia, 50, 39 –53.

Brown, J.H. (1995) Macroecology. University of Chicago Press, Chicago.

Brown, J.H. & Maurer, B.A. (1986) Body size, eco-logical dominance and Cope’s rule. Nature, 324, 248 – 250.

Brown, J.H. & Maurer, B.A. (1987) Evolution of spe-cies assemblages: effects of energetic constraints and species dynamics on the diversification of the North American avifauna. American Naturalist,

130, 1–17.

Brown, J.H. & Maurer, B.A. (1989) Macroecology: the division of food and space among species on continents. Science, 243, 1145 –1150.

Cade, B.S., Terrell, J.W. & Schroeder, R.L. (1999) Estimating effects of limiting factors with regres-sion quantiles. Ecology, 80, 311– 323.

Caswell, H. (1988) Theory and models in ecology: a different perspective. Ecological Modelling, 43, 33 – 44.

Chazdon, R.L., Colwell, R.K. & Denslow, J.S. (1999) Tropical tree richness and resource-based niches.

Science, 285, 1459.

Colwell, R.K. & Hurtt, G.C. (1994) Nonbiological gradients in species richness and a spurious Rap-oport effect. American Naturalist, 144, 570 –595.

(6)

342

N. J. Gotelli

Colwell, R.K. & Lees, D.L. (2000) The mid-domain effect: geometric constraints on the geography of species richness. Trends in Ecology and Evolution,

15, 70 –76.

Colwell, R.K. & Winkler, D.W. (1984) A null model for null models in biogeography. Ecological

Com-munities: Conceptual Issues and the Evidence (ed.

by D.R. Strong Jr, D. Simberloff, L.G. Abele and A.B. Thistle), pp. 344 –359. Princeton University Press, Princeton.

Connor, E.F. & Simberloff, D. (1979) The assembly of species communities: chance or competition?

Ecology, 60, 1132 –1140.

Connor, E.F. & Simberloff, D. (1983) Interspecific competition and species co-occurrence patterns on islands: null models and the evaluation of evidence.

Oikos, 41, 455– 465.

Connor, E.F. & Simberloff, D. (1986) Competition, scientific method, and null models in ecology.

American Scientist, 74, 155 –162.

Damuth, J. (1981) Population density and body size in mammals. Nature, 290, 699 –700.

Darwin, C. (1859) The Origin of Species by Means

of Natural Selection. Murray, London.

Diamond, J.M. (1975) Assembly of species commu-nities. Ecology and Evolution of Communities (ed. by M.L. Cody and J.M. Diamond), pp. 342 – 444. Harvard University Press, Cambridge.

Diamond, J.M. & Gilpin, M.E. (1982) Examination of the ‘null’ model of Connor and Simberloff for species co-occurrences on islands. Oecologia, 52, 64 –74.

Elton, C. (1946) Competition and the structure of ecological communities. Journal of Animal Ecology,

15, 54 – 68.

Enquist, B.J., Brown, J.H. & West, G.B. (1998) Allo-metric scaling of plant energetics and population density. Nature, 395, 163 –165.

Enquist, B.J., Jordan, M.A. & Brown, J.H. (1995) Connections between ecology, biogeography, and paleobiology: relationship between local abundance and geographic distribution in fossil and recent molluscs. Evolutionary Ecology, 9, 586 – 604. Enquist, B.J., West, G.B. & Charnov, E.L. (1999)

Allometric scaling of production and life-history variation in vascular plants. Nature, 401, 907 – 911. Fox, B.J. (1987) Species assembly and the evolution of community structure. Evolutionary Ecology, 1, 201– 213.

Fox, B.J. (1999) The genesis and development of guild assembly rules. Ecological Assembly Rules:

Perspectives, Advances, Retreats (ed. by E. Weiher

and P. Keddy), pp. 23 – 57. Cambridge University Press, Cambridge.

Fox, B.J. & Brown, J.H. (1993) Assembly rules for functional groups in North American desert rodent communities. Oikos, 67, 358 – 370.

Fox, B.J. & Kirkland, G.L. Jr (1992) An assembly rule for functional groups applied to North American

soricid communities. Journal of Mammology, 73, 491–503.

Garvey, J.E., Marschall, E.A. & Wright, R.A. (1998) From star charts to stoneflies: detecting relation-ships in continuous bivariate data. Ecology, 79, 442 – 447.

Gilpin, M.E. & Diamond, J.M. (1982) Factors contrib-uting to non-randomness in species co-occurrences on islands. Oecologia, 52, 75 – 84.

Gotelli, N.J. (1999) How do communities come together?

Science, 286, 1684 –1685.

Gotelli, N.J. (2000) Null model analysis of species co-occurrence patterns. Ecology, 81, 2606 –2621. Gotelli, N.J., Buckley, N.J. & Wiens, J.A. (1997)

Co-occurrence of Australian land birds: Diamond’s assembly rules revisited. Oikos, 80, 311–324. Gotelli, N.J. & Colwell, R.K. (in press) Quantifying

biodiversity: procedures and pitfalls in the measure-ment and comparison of species richness. Ecology Letters, in press.

Gotelli, N.J. & Entsminger, G.L. (2000) Ecosim: Null

Models Software for Ecology. Acquired Intelligence

& Kesey-Bear Inc., http://homepages.together.net/ ~gentsmin / ecosim.htm.

Gotelli, N.J. & Entsminger, G.L. (in press) Swap and fill algorithms in null model analysis: rethinking the Knight’s Tour. Oecologia, in press.

Gotelli, N.J. & Graves, G.R. (1996) Null Models in Eco-logy. Smithsonian Institution Press, Washington, DC. Gotelli, N.J. & McCabe, D.J. (in press) Species co-occurrence: a meta-analysis of Diamond’s (1975) assembly rules model. Ecology.

Graves, G.R. & Gotelli, N.J. (1983) Neotropical land-bridge avifaunas: new approaches to null hypotheses in biogeography. Oikos, 41, 322 –333.

Harvey, P.H., Colwell, R.K., Silvertown, J.W. & May, R.M. (1983) Null models in ecology. Annual

Review of Ecology and Systematics, 14, 189 –211.

Heck, K.L. Jr, van Belle, G. & Simberloff, D. (1975) Explicit calculation of the rarefaction diversity measurement and the determination of sufficient sample size. Ecology, 56, 1459 –1461.

Hilborn, R. & Mangel, M. (1997) The Ecological

Detective: Confronting Models with Data. Princeton

University Press, Princeton.

Hubbell, S.P., Foster, R.P., O’Brien, S.T., Harrms, K.E., Condit, R., Wechsler, B., de Wright, S.J. & Lao, S.L. (1999) Light-gap disturbances, recruitment limita-tion, and tree diversity in a neotropical forest.

Science, 283, 554 –557.

Hurlbert, S.H. (1971) The nonconcept of species divers-ity: a critique and alternative parameters. Ecology,

52, 577 –585.

Järvinen, O. (1982) Species-to-genus ratios in biogeo-graphy: a historical note. Journal of Biogeography,

9, 363 –370.

Kochmer, J.P. & Handel, S.N. (1986) Constraints and competition in the evolution of flowering phenology. Ecological Monographs, 56, 303 –325.

(7)

Null model analysis

343

Losos, J.B., Naeem, S. & Colwell, R.K. (1989) Hutch-insonian ratios and statistical power. Evolution,

43, 1820 –1826.

Maillefer, A. (1929) Le coefficient generique de P. Jacard et sa signification. Memoire Soc. Vaudoise Sc. Nat, 3, 113 –183.

Manly, B.F.J. (1991) Randomization and Monte Carlo

Methods in Biology. Chapman & Hall, London.

Manly, B.F.J. (1995) A note on the analysis of species co-occurrences. Ecology, 76, 1109 –1115. Maurer, B.A. (1999) Untangling Ecological

Complex-ity: the Macroecological Perspective. University of

Chicago Press, Chicago.

McCabe, D.J. & Gotelli, N.J. (2000) Effects of dis-turbance frequency, intensity, and area on assem-blages of stream invertebrates. Oecologia, 124, 270 –279.

Nitecki, M.H. & Hoffman, A. (1987 ) Neutral Models

in Biology. Oxford University Press, Oxford.

Penev, L. (1997) Concrete biotas — a neglected concept in biogeography? Global Ecology and Biogeography Letters, 6, 91– 96.

Popper, K.R. (1959) The Logic of Scientific Discovery

(revised, 1965 edition). Basic Books, New York. Preston, F.W. (1962) The canonical distribution of

commonness and rarity. Ecology, 43, 185 – 215. Roberts, A. & Stone, L. (1990) Island-sharing by

archipelago species. Oecologia, 83, 560 –567. Roughgarden, J. (1983) Competition and theory in

community ecology. American Naturalist, 122, 583 – 601.

Roxburgh, S.H. & Chesson, P. (1998) A new method for detecting species associations with spatially autocorrelated data. Ecology, 79, 2180 –2192. Roxburgh, S.H. & Matsuki, M. (1999) The statistical

validation of null models used in spatial associ-ation analysis. Oikos, 85, 68 –78.

Sanders, H. (1968) Marine benthic diversity: a com-parative study. American Naturalist, 102, 243 –282. Sanderson, J.G., Moulton, M.P. & Selfridge, R.G. (1998) Null matrices and the analysis of species co-occurrences. Oecologia, 116, 275 –283. Scharf, F.S., Juanes, F. & Sutherland, M. (1998)

Inferring ecological relationships from the edges of scatter diagrams: comparisons of regression techniques. Ecology, 79, 448 – 460.

Shenk, T.M., White, G.C. & Burnham, K.P. (1998) Sampling-variance effects on detecting density dependence from temporal trends in natural popu-lations. Ecological Monographs, 68, 421– 444. Simberloff, D. (1970) Taxonomic diversity of island

biotas. Evolution, 24, 23– 47.

Stevens, M.H.H. & Carson, W.P. (1999) Plant density determines species richness along an experimental fertility gradient. Ecology, 80, 455– 465.

Stone, L. & Roberts, A. (1990) The checkerboard score and species distributions. Oecologia, 85, 74 – 79.

Strong, D.R. Jr, Simberloff, D., Abele, L.G. & Thistle, A.B. (eds) (1984) Ecological Communities:

Conceptual Issues and the Evidence. Princeton

University Press, Princeton.

Thomson, J.D., Weiblen, G., Thomson, B.A., Alfaro, S. & Legendre, P. (1996) Untangling multiple factors in spatial distributions: lillies, gophers, and rocks.

Ecology, 77, 1698 –1715.

Tipper, J.C. (1979) Rarefaction and rarefiction — the use and abuse of a method in paleoecology. Paleo-biology, 5, 423 – 434.

Van Valen, L. (1973) Body size and numbers of plants and animals. Evolution, 27, 27 –35.

Weiher, E. & Keddy, P. (eds) (1999) Ecological Assembly

Rules: Perspectives, Advances, Retreats. Cambridge

University Press, Cambridge.

West, G.B., Brown, J.H. & Enquist, B.J. (1997) A general model for the origin of allometric scaling laws in biology. Science, 276, 122 –126.

West, G.B., Brown, J.H. & Enquist, B.J. (1999a) A general model for the structure and allometry of plant vascular systems. Nature, 400, 664 – 667. West, G.B., Brown, J.H. & Enquist, B.J. (1999b) The

fourth dimension of life: fractal geometry and allo-metric scaling of organisms. Science, 284, 1677–1679. Wiens, J.A. (1977) On competition and variable environments. American Scientist, 65, 590 –598. Wiens, J.A. (1989) The Ecology of Bird Communities,

Vol. 1. Foundations and Patterns (ed. by R.S.K. Barnes, H.J.B. Birks, E.F. Connor and R.T. Paine), pp. 1– 539. Cambridge Studies in Ecology, Cambridge University Press, Cambridge.

Williams, C.B. (1947) The generic relations of species in small ecological communities. Journal of Animal

Ecology, 16, 11–18.

Williams, C.B. (1964) Patterns in the Balance of Nature. Academic Press, New York.

Wilson, J.B. (1987) Methods for detecting non-randomness in species co-occurrences: a contribu-tion. Oecologia, 73, 579 –582.

Wilson, J.B. (1995) Null models for assembly rules: the Jack Horner effect is more insidious than the Narcissus effect. Oikos, 72, 139 –144.

Wilson, J.B. (1999) Assembly rules in plant com-munities. Ecological Assembly Rules: Perspectives,

Advances, Retreats (ed. by E. Weiher & P. Keddy),

pp. 130 –164. Cambridge University Press, Cambridge. Wilson, J.B. & Gitay, H. (1995) Limitations to spe-cies coexistence: evidence for competition from field observations, using a patch model. Journal of

Vegetation Science, 6, 369–376.

Winemiller, K.O. & Pianka, E.R. (1990) Organiza-tion in natural assemblages of desert lizards and tropical fishes. Ecological Monographs, 60, 27–55. Zaman, A. & Simberloff, D. (in press) Random binary matrices in biogeographical ecology — instituting a good neighbor policy. Environmental and Eco-logical Statistics.

Referencias

Documento similar

In Section 2 we consider abstract evolution equations with memory terms and prove that the null and memory-type null controllability properties are equivalent to certain

We construct a one-loop effective metric describing the evaporation phase of a Schwarzschild black hole in a spherically symmetric null- dust model.. This is achieved by quantising

The interaction effects of soil type, altitude, and type of forest stand on the total soil organic carbon stock were confirmed by the general linear model (GLM) analysis (Table

The current study adds to this body of research by examining the number and type of self-assessment strategies and criteria among higher education students in a ran- domized

In Figure 72, we show a simple model of an escalation process in response operations based on data form both case studies, Elbe Flood and Schiphol Fire.. The model is an

We also examine how well the null-test can be reconstructed by future data by creating mock catalogs based on the cosmological constant model, a model with strong dark

van Bilsen et al., “Long-term severe diabetes only leads to mild cardiac diastolic dysfunction in Zucker diabetic fatty rats,” European Journal of Heart Failure, vol.. Ares-Carrasco

Subsequently and in a deeper way, Valdés (1998) highlights the main lines of research studied between personality and sport: 1) analysis of the selection of a specific type